It’s time to teach peer review as a core research skill

The rapidly growing volume of academic papers is outpacing the scientific community’s ability to critically evaluate them. This makes it increasingly important to equip the next generation of researchers with the skills that they need to become peer reviewers.

But formal curricula rarely include these skills. Universities invest years in teaching students how to design experiments and analyse data, but leave peer-review training mostly to informal mentorship.

So, how do early-career researchers — especially students and postdocs — learn to develop a critical eye?

Preprints offer a valuable opportunity for trainees to practise reviewing manuscripts outside the formal journal-review process, and several initiatives have been established with this in mind.

PREreview, for example, is a platform on which researchers can openly review and discuss preprints. And Preprint Club, a cross-institutional journal club run by an international group of universities, allows early-career researchers to select manuscripts for discussion so that they can learn from experience alongside their peers, more-senior researchers and faculty mentors.

Inspired by these efforts, we worked with the Bakar ImmunoX Initiative, a research programme at the University of California, San Francisco (UCSF), to launch our own preprint journal club, called ImmunoX Preprint JC, over the 2025–26 academic year.

We think that the journal club’s greatest value is in helping early-career researchers to build confidence and expertise in peer review. As a result, they are included in every level of conversation, from shaping how the club is run and how preprints are scored to leading discussions during journal-club sessions.

We also feel that other institutions could emulate this model, which combines hands-on experience, mentoring and relationship building, to support the development of their own early-career researchers.

Group discussion

Each monthly ImmunoX Preprint JC session follows the steps below. A postdoc starts by selecting an immunology-related preprint posted on the bioRxiv server. They choose an ImmunoX faculty ‘coach’, who helps them to assess the preprint on the basis of its scientific quality, impact and novelty. The manuscript and scoring rubric are shared with the wider ImmunoX community before the meeting.

At the meeting, the postdoc leads a 20-minute presentation discussing the preprint’s strengths and proposing experiments to improve it. Attendees then have a five- to ten-minute discussion about the preprint’s strengths and weaknesses; this is followed by anonymous scoring of the manuscript in terms of its scientific quality, impact and novelty.

Each manuscript’s publication status is tracked over several weeks to assess how well the ImmunoX community evaluated it. We plan to have the presenter of the top-scoring preprint over a six-month period write up a commentary with their coach. The piece would then be shared with the ImmunoX community and general public online.

Tracking progress

A key goal for the ImmunoX Preprint JC is to gauge whether the community-based assessment of a preprint reflects whether it is eventually accepted by a journal. Such data can establish a benchmark for how well our community — especially trainees — evaluates the scientific merits of a preprint.

We consulted with the Immunology Preprint Club at the Icahn School of Medicine at Mount Sinai in New York City to compare scores given by both groups to the same preprints.

We also compared the preprints’ real-life performance with our scores. Although most of the preprints we have so far reviewed remain unpublished, our top-scoring paper was subsequently published in Cell¹ — a leading life-sciences journal — suggesting that community-based assessments can, at least in some cases, identify work that will go on to meet the standards of a high-profile journal. Our findings from the 2025–26 academic year can be viewed in more detail on the ImmunoX website.

Assessing our own data allows us not only to track occasions we’ve correctly identified high-quality work, but also to pick up on potential biases. For example, we noticed that manuscripts whose senior authors were highly cited tended to receive higher scores in our journal-club sessions. This could mean that our community reliably identified rigorous science, or it might reflect an unconscious bias correlating reputation with scientific quality.

Learning to notice this will help trainees to develop more self-awareness, which is crucial for an unbiased peer-review process. In the future, we plan to withhold manuscript author names until the end of a scoring session.